Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial...

10
ANNALS OF FAMILY MEDICINE WWW.ANNFAMMED.ORG VOL. 11, NO. 1 JANUARY/FEBRUARY 2013 43 Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of Care Among Black and Latino Patients ABSTRACT PURPOSE We investigated whether clinicians’ explicit and implicit ethnic/racial bias is related to black and Latino patients’ perceptions of their care in estab- lished clinical relationships. METHODS We administered a telephone survey to 2,908 patients, stratified by ethnicity/race, and randomly selected from the patient panels of 134 clinicians who had previously completed tests of explicit and implicit ethnic/racial bias. Patients completed the Primary Care Assessment Survey, which addressed their cli- nicians’ interpersonal treatment, communication, trust, and contextual knowledge. We created a composite measure of patient-centered care from the 4 subscales. RESULTS Levels of explicit bias were low among clinicians and unrelated to patients’ perceptions. Levels of implicit bias varied among clinicians, and those with greater implicit bias were rated lower in patient-centered care by their black patients as compared with a reference group of white patients (P = .04). Latino patients gave the clinicians lower ratings than did other groups (P <.0001), and this did not depend on the clinicians’ implicit bias (P = .98). CONCLUSIONS This is among the first studies to investigate clinicians’ implicit bias and communication processes in ongoing clinical relationships. Our findings suggest that clinicians’ implicit bias may jeopardize their clinical relationships with black patients, which could have negative effects on other care processes. As such, this finding supports the Institute of Medicine’s suggestion that clinician bias may contribute to health disparities. Latinos’ overall greater concerns about their clinicians appear to be based on aspects of care other than clinician bias. Ann Fam Med 2013;11:43-52. doi:10.1370/afm.1442. INTRODUCTION P rimary care clinicians serve as the cornerstone of the health care system and are required to possess many skills. Patient-centeredness is 1 of 6 key dimensions of high-quality health care, 1 and if clinicians are to provide such care, they must be able to engage patients in a collab- orative partnership. The importance of the clinician-patient relationship is underscored by demonstrated links between the quality of the relationship and a number of processes and outcomes of care, including patients’ adherence to medi- cal advice, 2-6 decision to remain with a clinician, 7 satisfaction with care, 3 and clinical outcomes of care. 3,8,9 Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care. 10-17 In addition to cultural and language barriers, there have been long-standing concerns that clinician bias may contribute to lower- quality clinical relationships. 18,19 A 2003 report 18 by the Institute of Medicine noted, “It is likely that the vast majority [of clinicians] endorse egalitarian and non-racist attitudes.” Irene V. Blair, PhD 1 John F. Steiner, MD, MPH 2 Diane L. Fairclough, DrPH 3 Rebecca Hanratty, MD 4 David W. Price, MD 2 Holen K. Hirsh, PhD 1 Leslie A. Wright, MA 2 Michael Bronsert, PhD 3 Elhum Karimkhani, MPH 5 David J. Magid, MD 2 Edward P. Havranek, MD 5 1 Department of Psychology and Neuro- science, University of Colorado Boulder, Boulder, Colorado 2 Institute for Health Research, Kaiser Permanente Colorado, Denver, Colorado 3 Colorado Health Outcomes Center, University of Colorado Denver, Denver, Colorado 4 Division of Internal Medicine, Denver Health, Denver, Colorado 5 Division of Cardiology, Denver Health, Denver, Colorado Conflicts of interest: authors report none. CORRESPONDING AUTHOR Irene V. Blair, PhD University of Colorado Boulder Department of Psychology and Neuroscience Boulder, CO 80309-0345 [email protected]

Transcript of Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial...

Page 1: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

43

Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of Care Among Black and Latino Patients

ABSTRACTPURPOSE We investigated whether clinicians’ explicit and implicit ethnic/racial bias is related to black and Latino patients’ perceptions of their care in estab-lished clinical relationships.

METHODS We administered a telephone survey to 2,908 patients, stratifi ed by ethnicity/race, and randomly selected from the patient panels of 134 clinicians who had previously completed tests of explicit and implicit ethnic/racial bias. Patients completed the Primary Care Assessment Survey, which addressed their cli-nicians’ interpersonal treatment, communication, trust, and contextual knowledge. We created a composite measure of patient-centered care from the 4 subscales.

RESULTS Levels of explicit bias were low among clinicians and unrelated to patients’ perceptions. Levels of implicit bias varied among clinicians, and those with greater implicit bias were rated lower in patient-centered care by their black patients as compared with a reference group of white patients (P = .04). Latino patients gave the clinicians lower ratings than did other groups (P <.0001), and this did not depend on the clinicians’ implicit bias (P = .98).

CONCLUSIONS This is among the fi rst studies to investigate clinicians’ implicit bias and communication processes in ongoing clinical relationships. Our fi ndings suggest that clinicians’ implicit bias may jeopardize their clinical relationships with black patients, which could have negative effects on other care processes. As such, this fi nding supports the Institute of Medicine’s suggestion that clinician bias may contribute to health disparities. Latinos’ overall greater concerns about their clinicians appear to be based on aspects of care other than clinician bias.

Ann Fam Med 2013;11:43-52. doi:10.1370/afm.1442.

INTRODUCTION

Primary care clinicians serve as the cornerstone of the health care

system and are required to possess many skills. Patient-centeredness

is 1 of 6 key dimensions of high-quality health care,1 and if clinicians

are to provide such care, they must be able to engage patients in a collab-

orative partnership.

The importance of the clinician-patient relationship is underscored by

demonstrated links between the quality of the relationship and a number

of processes and outcomes of care, including patients’ adherence to medi-

cal advice,2-6 decision to remain with a clinician,7 satisfaction with care,3

and clinical outcomes of care.3,8,9

Ethnic/racial minorities appear to be at a disadvantage in this aspect of

health care.10-17 In addition to cultural and language barriers, there have

been long-standing concerns that clinician bias may contribute to lower-

quality clinical relationships.18,19

A 2003 report18 by the Institute of Medicine noted, “It is likely that the

vast majority [of clinicians] endorse egalitarian and non-racist attitudes.”

Irene V. Blair, PhD1

John F. Steiner, MD, MPH2

Diane L. Fairclough, DrPH3

Rebecca Hanratty, MD4

David W. Price, MD2

Holen K. Hirsh, PhD1

Leslie A. Wright, MA2

Michael Bronsert, PhD3

Elhum Karimkhani, MPH5

David J. Magid, MD2

Edward P. Havranek, MD5

1Department of Psychology and Neuro-

science, University of Colorado Boulder,

Boulder, Colorado

2Institute for Health Research, Kaiser

Permanente Colorado, Denver, Colorado

3Colorado Health Outcomes Center,

University of Colorado Denver, Denver,

Colorado

4Division of Internal Medicine, Denver

Health, Denver, Colorado

5Division of Cardiology, Denver Health,

Denver, Colorado

Confl icts of interest: authors report none.

CORRESPONDING AUTHOR

Irene V. Blair, PhD

University of Colorado Boulder

Department of Psychology and

Neuroscience

Boulder, CO 80309-0345

[email protected]

Page 2: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

44

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

But also, “[there is] strong but circumstantial evidence

for the role of bias, stereotyping, prejudice, and clini-

cal uncertainty” in the genesis of health disparities.

The observation that clinicians are unlikely to directly

express ethnic/racial bias yet may still deliver care that

is infl uenced by unrecognized bias is consistent with

research in social psychology that demonstrates that

bias exists on explicit and implicit levels. Whereas

explicit bias is overt and freely expressed, implicit

bias may not be consciously acknowledged and oper-

ates in more subtle ways.20-22 For example, a clinician

with implicit bias may unconsciously exhibit negative

behavior or poor communication with a black patient,

as has been shown in laboratory research.21,23-25 In

addition to reducing the patient’s comfort and trust,

such actions may impede the fl ow of information, lead

to shorter interviews, and reduce the patient’s under-

standing of and resolve to follow medical advice.2-6,13

Few studies have directly investigated whether clini-

cian bias is related to communication, interpersonal

treatment, and trust in ongoing clinical relationships.

In a prior study,26 we found that two-thirds of par-

ticipating primary care clinicians showed some implicit

ethnic/racial bias that favored whites, even as they

rejected explicit expressions of bias. In the current

study, we contacted a large sample of patients of those

clinicians and asked them to rate the degree to which

their clinicians were patient-centered during their

interactions. We then examined those ratings as a func-

tion of the patients’ ethnicity/race and the clinicians’

implicit and explicit ethnic/racial bias. We hypoth-

esized that clinicians with higher levels of implicit bias

would be rated less favorably by their minority patients

than clinicians with lower levels of implicit bias.

METHODS Study Population and Data Collection: Clinicians We derived data about the implicit bias of clinicians

from a prior study26 in which adult primary care clini-

cians from 3 health care organizations were invited to

complete measures of their implicit and explicit ethnic/

racial attitudes (N = 210, 60% participation rate). Data

from that study were included in the present study for

2 organizations: Denver Health and Kaiser Permanente

Colorado. Denver Health is an integrated safety-net

health care system and is nationally recognized for its

model of care to underserved, indigent, and minor-

ity patients.27,28 Denver Health community clinics see

more than 100,000 unique patients each year (15%

black, 60% Latino, 16% white, and 9% other). Kaiser

Permanente Colorado is a closed-panel, group-model,

not-for-profi t health maintenance organization with

approximately 480,000 members in the Denver area

(5% black, 17% Latino, 74% white, and 4% other). The

institutional review board for each institution approved

the study design and procedures.

Implicit Bias

We measured implicit bias against blacks and against

Latinos with 2 Implicit Association Tests (IATs).29

The IAT measures implicit bias29-32 by the speed with

which a person can respond to a group and positive vs

negative words. Implicit bias is shown, for example, if

the person is signifi cantly faster when black faces and

negative words require the same response while white

faces and good words require another response, com-

pared with the reverse pairing. The larger this perfor-

mance difference, the stronger the implicit bias for that

person (demonstration is available at https://implicit.

harvard.edu). The IAT has been widely used, and its

psychometric properties and methodologic strengths

and limitations have been extensively reviewed.30-34

The 2 IATs completed by the clinicians in this study

were validated in previous research to measure implicit

bias against blacks compared with whites and against

Latinos compared with whites.26,35 Possible scores

ranged from –2 to +2, with negative scores indicat-

ing bias against whites, positive scores indicating bias

against blacks or Latinos, and 0 indicating no bias.

Explicit Bias

The clinicians were asked to indicate their explicit atti-

tudes toward blacks, Latinos, and whites on 2 standard

measures36,37: Feeling Thermometers (with possible

scores of 0 to 100 for “cool” to “warm” feelings) and a

set of semantic differential scales (7-point trait ratings

of “hard-working to lazy,” “wise to foolish,” and “coop-

erative to hostile”).

Study Population and Data Collection: PatientsWe obtained primary data from patients in a broader

study on hypertension care; thus, all patients had

diagnosed hypertension. Patients were included in the

sampling frame for this study if they received regular

care from a participating clinician and their ethnic-

ity/race of record was black, Latino, or white. The

patients were stratifi ed by clinician and ethnicity/race,

and then randomly selected within each stratum up

to a maximum of 12 patients. This initial screening

produced 7,437 patients, of whom 1,308 were subse-

quently determined to be ineligible for the study (1,055

had incorrect contact information, 210 did not con-

fi rm their primary care clinician, and 43 self-identifi ed

their ethnicity/race as other than any of the 3 groups

included in this study).

A professional survey company attempted to call

the 6,129 eligible patients by telephone and adminis-

Page 3: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

45

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

ter the questionnaire in either English or Spanish. If

telephone contact could not be made, a written ques-

tionnaire in both English and Spanish was sent to the

patient’s last known address. All participants were sent

a $10 gift card.

Patient Survey

Four subscales were administered from the well-

validated Primary Care Assessment Survey38 (PCAS):

interpersonal treatment, communication, trust, and

contextual knowledge. Example items include rating the

clinician on “caring and concern for you” (interpersonal

treatment), “you leave your doctor’s offi ce with unan-

swered questions” (communication), “my doctor some-

times pretends to know things when really not sure”

(trust), and “knowledge about you as a person (your val-

ues and beliefs)” (contextual knowledge). Each subscale

is scored from 0 to 100, with higher scores indicating

a higher level of the attribute; the 4 subscales were

averaged to create a composite measure of patient-cen-

teredness. Additional survey questions assessed patients’

sociodemographic characteristics. The survey was con-

ducted from mid-2010 to early 2011.

Identifi cation of Patients’ Primary Care Clinician

A clinician was identifi ed for a patient only if the

patient (1) saw that clinician for a majority of primary

care visits in 3 years, (2) visited that clinician at least

3 times in that period, and (3) confi rmed on the ques-

tionnaire that he/she received regular care from that

clinician. The patient was also asked how many years

he/she had been going to that clinician.

Ethnicity/Race

Patients were fi rst screened by the ethnicity/race

recorded in their medical fi les to include only black,

Latino, or white patients. Those who were subse-

quently contacted were asked to identify their ethnic-

ity/race (with results showing 90% agreement with

records), and their self-identifi ed ethnicity/race was

used for analysis.

Statistical Analyses We evaluated differences in patients’ demographics

using the Pearson χ2 test for categorical variables, the

nonparametric Kruskal-Wallis test for ordered categor-

ical variables, and analysis of variance for continuous

variables. The primary dependent variables were the

patients’ responses to the 4 PCAS subscales and the

composite measure of patient-centeredness, with the

patients’ ethnicity/race and each measure of their clini-

cians’ ethnic/racial bias used as predictors. The specifi c

effect of interest was the degree to which black or

Latino patients differed from white patients in their

ratings of the clinicians, and critically, the extent to

which those differences were themselves predicted by

the clinicians’ ethnic/racial bias (analyzed as continu-

ous variables). Measures of clinician bias against blacks

were used in predicting black patients’ ratings; mea-

sures of clinician bias against Latinos were used in pre-

dicting Latino patients’ ratings. Because patients were

nested under clinicians, who were themselves nested

under clinics, the data were analyzed using hierarchi-

cal linear modeling or mixed effects models. White

patients were always used as the reference group.

We also considered a nonlinear relationship between

patient’s survey ratings and clinician bias using a simple

linear spline with 5 knots.

Supplementary analyses were conducted to exam-

ine additional patient background characteristics

(sex, age, socioeconomic status, and, for Latinos only,

Spanish and English language profi ciency) as statisti-

cal controls and to assess subgroup differences. Age

was analyzed in decades, and socioeconomic status

was analyzed as a dichotomous variable in terms of

education (high school or less vs at least some college).

Reported income was not used for socioeconomic

status because of missing data. Spanish and English

language profi ciency was analyzed as a dichotomous

variable: patients were coded as having greater Spanish

than English profi ciency if they chose to complete the

survey in Spanish or they reported that they were fl u-

ent in Spanish but less than fl uent in English; all other

patients were assigned to the alternate category.

RESULTS From the original group of 210 clinicians who had

completed the measures of ethnic/racial bias,26 134

(64%) met this study’s inclusion criteria. These clini-

cians’ characteristics—54% female, 75% white, and

50% with more than 10 years of clinical experience—

were nearly identical to those previously reported

for the full clinician sample.26 As in the full sample,

approximately two-thirds of the clinicians had implicit

bias against blacks (43% moderate to strong) and Lati-

nos (51% moderate to strong), while reporting very

little explicit bias against either group.

Of the 6,129 patients in the recruitment pool, 2,908

(47%) completed the survey questionnaire. Reasons for

nonparticipation varied: 1,878 were unreachable, 558

were unable (eg, because they had died or had long-

term disability), 780 declined, and 5 did not answer

enough questions. Characteristics of the participating

and nonparticipating patients are shown in Table 1.

Among participating patients, all 3 ethnic/racial

groups had well-established relationships with their

clinicians: two-thirds of each group saw their named

Page 4: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

46

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

clinician for at least 70% of their primary care visits,

the average clinical relationship had been ongoing for

more than 3 years, and there had been an average of

more than 7 visits with the clinician in 3 years.

Primary OutcomesPatients’ Ratings of Clinicians’ Patient-Centeredness

Patients in all 3 groups evaluated their clinicians favor-

ably on the measures of patient-centered care (Table

2), similar to what has been found in previous large-

scale studies.3,7,38-42 Compared with white patients,

black patients gave mostly equivalent ratings to the

clinicians (composite scale difference, P = .84), whereas

Latino patients gave comparatively lower ratings (com-

posite scale difference, P <.0001).

Patients’ Ratings as a Function of Clinicians’

Implicit Bias

There were consistent associations between clinicians’

implicit bias and their black patients’ evaluations of

them: the stronger the clinicians’ implicit preference for

whites over blacks, the lower their black patients rated

Table 1. Characteristics of Participating and Nonparticipating Patients

Characteristic

Recruitment Sample Final Patient Sample

by Ethnicity/Race (N = 2,908)

Nonparticipants (n = 3,221)

Participants(n = 2,908)

Black(n = 612)

Latino(n = 859)

White(n = 1,437)

Female, No. (%)a,b 1,690 (52) 1,694 (58) 369 (60) 539 (63) 786 (55)

Age, No. (%)a,b

18-35 y 108 (3) 67 (2) 25 (4) 29 (3) 13 (1)

36-55 y 1,115 (36) 973 (33) 234 (38) 314 (37) 425 (30)

≥56 y 1,958 (61) 1,868 (64) 353 (58) 516 (60) 999 (70)

Ethnicity/race, No. (%)a

Black 560 (17) 612 (21) – – –

Latino 1,146 (36) 859 (30) – – –

White 1,515 (47) 1,437 (49) – – –

Education, No. (%)b

High school not completed – 529 (18) 91 (15) 355 (42) 83 (5)

High school diploma or GED – 773 (27) 185 (30) 237 (28) 351 (24)

1-3 y college – 866 (30) 215 (35) 167 (19) 484 (34)

≥4 y college – 714 (24) 118 (20) 88 (11) 508 (35)

Unknown – 26 (1) 3 (<1) 12 (1) 11 (1)

Household income, No. (%)b

≤$15,000 – 1,026 (35) 274 (45) 396 (46) 356 (25)

$16,000-$35,000 – 555 (19) 133 (22) 184 (21) 238 (17)

$36,000-$55,000 – 408 (14) 68 (11) 88 (10) 252 (18)

≥$56,000 – 717 (25) 100 (16) 124 (14) 493 (34)

Unknown – 202 (7) 37 (6) 67 (8) 98 (7)

Language profi ciency,c No. (%)b

Spanish > English – 268 (9) 5 (1) 260 (30) 3 (<1)

Alternate category – 2,640 (91) 607 (99) 599 (70) 1,434 (>99)

Proportion of primary care visits with clinician, No. (%)b

.50-.59 520 (16) 446 (15) 85 (14) 137 (16) 224 (16)

.60-.69 530 (16) 471 (16) 84 (14) 133 (15) 254 (18)

.70-.79 552 (17) 519 (18) 99 (16) 161 (19) 259 (18)

.80-.89 644 (20) 594 (20) 144 (24) 182 (21) 268 (19)

.90-1.0 975 (30) 878 (30) 200 (33) 246 (29) 432 (30)

Visits with clinician in 3 y, mean No. (SD)a,b

7.38 (5.35)

7.81 (5.70)

8.16 (6.05)

8.51 (5.91)

7.23 (5.36)

Years with clinician, mean No. (SD)

– 3.40 (1.07)

3.35 (1.09)

3.37 (1.16)

3.45 (1.00)

GED = general equivalency degree.

a Participants and nonparticipants differ, P <.05. b Ethnic/racial groups differ, P <.05.c Greater profi ciency in Spanish than English was assigned if (1) patients completed the questionnaire in Spanish instead of English, or (2) patients reported on the questionnaire that they were fl uent in Spanish and less than fl uent in English.

Page 5: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

47

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

them. This negative association

was seen to varying degrees on all

4 subscales (Figure 1) and on the

composite scale (t = 2.05, P = .04);

Table 3 shows model estimates

and Figure 2 shows predicted

composite scale values for indi-

vidual clinicians. As a concrete

example, black patients rated

clinicians who scored 1.0 on the

IAT (strong bias) approximately

6 points lower on interpersonal

treatment than clinicians who

scored 0 on the IAT (no bias).

In contrast, there was no asso-

ciation between Latino patients’

ratings and their clinicians’

implicit ethnic/racial bias on any

of the 4 subscales or the compos-

ite scale (t = 0.03, P = .98). Tests

of nonlinearity with knots at IAT

Table 2. PCAS Scores by Patients’ Ethnicity/Race

Scale (α) and Description

Score, Mean (SD)

Black Latino White

Subscalea

Interpersonal treatment (α = .94); 5 items on the clinician’s patience, friendliness, caring, respect, and time spent with the patient

84 (19) 81b (19) 86 (18)

Communication (α = .93); 6 items on the thorough-ness of the clinician’s questions, attention to the patient, clarity of explanations and instructions, and help in making decisions about care

84 (18) 80b (19) 84 (17)

Trust (α = .85); 8 items on the clinician’s integrity and role as the patient’s agent in the system

79b (16) 76b (15) 82 (15)

Contextual knowledge (α = .90); 5 items on the clinician’s knowledge of the patient’s medical history, life responsibilities, principal health con-cerns, and values and beliefs

75 (19) 73 (20) 74 (20)

Composite (α = .93); average of all 4 subscales weighted equally

80 (16) 78b (17) 82 (16)

PCAS = Primary Care Assessment Survey.

Note: α is a measure of internal reliability.a Each subscale is scored from 0 to 100, with higher scores indicating a higher level of the attribute.b Score is less than that for white patients, P <.01.

Figure 1. Predicted ratings of clinicians as a function of their implicit bias (IAT) score and their patients’ ethnicity/race. White patients always served as the reference group (data not shown).

IAT = Implicit Association Test.

Pred

icte

d R

atin

gs

100 100

95

90

85

80

75

70

65

0

95

90

85

80

75

70

65

0

–1.0 1.0 –1.0 1.0Strongly

Prefer BlacksStrongly

Prefer WhitesStrongly

Prefer LatinosStrongly

Prefer Whites

Clinician Implicit Bias Score on Black:White IAT Clinician Implicit Bias Score on Latino:White IAT

Interpersonal Treatment Communication Trust Contextual Knowledge

Black Patients Latino Patients

Page 6: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

48

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

scores of –0.65, –0.35, 0, 0.35, and 0.65, showed that

all associations (or lack thereof) were similar in magni-

tude across the range of bias scores (data not shown).

Patients’ Ratings as a Function of Clinicians’

Explicit Bias

Neither the thermometer nor the trait rating measures

of clinicians’ explicit ethnic/racial bias was associated

with patients’ ratings of patient-centered care, for black

patients (composite scale P = .13 and .23) or for Latino

patients (composite scale P = .23 and .16).

Subgroup Analyses

For black patients, only age moderated the association

between patients’ ratings and clinicians’ implicit bias, so

that the negative association was signifi cantly stronger

Table 3. Effects of Ethnic/Racial Group and Interactions Between Group and Clinicians’ Implicit Bias on Patients’ PCAS Ratings of the Clinicians

Predictor

PCAS Score by Scale, Estimate (SE)

Interpersonal Treatment Communication Trust

Contextual Knowledge Composite

Intercept (average rating by white patients) 85.69 (0.57) 84.45 (0.56) 81.88 (0.47) 74.36 (0.62) 81.62 (0.52)

Black patients

Group (black vs white) 0.24 (1.11) 0.64 (1.10) –2.25 (0.93)a 2.42 (1.22)a 0.20 (1.00)

Group × clinicians’ implicit bias –5.81 (2.52)a –4.31 (2.47)b –2.65 (2.09) –5.58 (2.73)a –4.61 (2.25)a

Latino patients

Group (Latino vs white) –4.30 (0.97)c –3.93 (0.96)c –5.85 (0.81)c –1.31 (1.06) –3.86 (0.87)c

Group × clinicians’ implicit bias –0.58 (1.71) –0.13 (1.68) 0.85 (1.42) –0.19 (1.86) –0.04 (1.53)

PCAS = Primary Care Assessment Survey; SE = standard error.

a P <.05.b P <.09.c P <.0001.

Figure 2. Predicted composite scale ratings by black, Latino, and white (reference) patients for individual clinicians with specifi c IAT scores.

IAT = Implicit Association Test.

Note: The lines show the overall (unconditional) estimate of the relation between clinician implicit bias scores and predicted patient ratings. The symbols show the indi-vidual (conditional) estimates for each clinician with a specifi c IAT score by each ethnic/racial patient group.

100

88

84

80

76

72

0

88

84

80

76

72

0

Pred

icte

d C

ompos

ite

Scal

e Rat

ing

100

–1.0Strongly

Prefer Blacks

1.0Strongly

Prefer Whites

–1.0Strongly

Prefer Latinos

1.0Strongly

Prefer Whites

Clinician Implicit Bias Score on Black:White IAT Clinician Implicit Bias Score on Latino:White IAT

Black Latino White (reference)

88

84

80

76

72

100

Black Latino White (reference)

Page 7: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

49

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

for younger than older patients; the race-by-age inter-

action was signifi cant or nearly so for interpersonal

treatment (P = .06), communication (P = .01), trust

(P = .01), and contextual knowledge (P = .07), as well

as for the composite score (P = .02) (Figure 3). As an

example, the model showed that among blacks aged 40

years, clinicians with an IAT score of 1.0 were rated 12

points lower on communication than clinicians with a

score of 0; among blacks aged 60 years, that difference

was only 2 points. None of the background character-

istics, including language, altered the primary fi ndings

for Latino patients (data not shown).

DISCUSSION Our data show that clinicians’ implicit ethnic/racial

bias is related to the quality of clinical relationships

for some patients: clinicians with greater implicit bias

against blacks were consistently evaluated as provid-

ing less patient-centered care by their black patients

than were clinicians with little or no such implicit bias.

We did not assess health outcomes in this study, but

prior research has shown that patients who evaluate

their clinicians more positively on these same mea-

sures of patient-centeredness are more satisfi ed with

their care,3 are more likely to adhere to treatment

and follow-up with their clinician,3-5,7 and have better

health outcomes.3,9

Although Latino patients generally gave their cli-

nicians lower ratings than did other patient groups,

these ratings were unrelated to the clinicians’ ethnic/

racial bias. Even subgroups shown previously to have

greater concerns with clinical interactions (eg, Spanish-

speaking Latinos43-45) did not provide lower evaluations

to more-biased clinicians. This is the fi rst study to

investigate the perceptions of Latino patients in rela-

tion to clinician bias, and the difference in fi ndings for

this group requires further investigation.

Only 2 prior studies46,47 have examined the link

between clinicians’ implicit bias and patients’ per-

ceptions, both with small samples of clinicians and

patients. One study46 found that black patients gave

lower ratings to clinicians having greater implicit race

bias, but only if the clinicians also reported very low

levels of explicit bias. The other study47 found more

consistent associations between black patients’ ratings

Figure 3. Predicted ratings of clinicians by younger and older black patients, as a function of clinicians’ implicit bias score on the Black:White IAT .

IAT = Implicit Association Test.

Interpersonal Treatment Communication Trust Contextual Knowledge

Pred

icte

d R

atin

gs

100

95

90

85

80

75

70

60

0

65

100

95

90

85

80

75

70

60

0

65

–1.0 1.0 –1.0 1.0Strongly

Prefer BlacksStrongly

Prefer WhitesStrongly

Prefer BlacksStrongly

Prefer Whites

Clinician Implicit Bias Score Clinician Implicit Bias Score

Patients Centered at 40 Years Old Patients Centered at 60 Years Old

Page 8: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

50

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

and clinicians’ implicit bias; however, even at higher

levels of implicit race bias, black patients in that study

tended to rate the clinicians more positively than did

the white patients, complicating the study conclusions.

The current fi ndings extend previous work on

patient perceptions in important ways. This study was

conducted with a robust sample of experienced clini-

cians from 2 different organizations, a large sample of

patients in ongoing clinical relationships with those

clinicians, validated measures of patient-centeredness

(associated in prior research with patient satisfaction,

adherence, and health outcomes), and a nested study

design that included a large sample of white patients to

more closely pinpoint ethnic/racial differences among

patients who see the same clinician.

Three additional studies48-51 have examined the

potential for clinicians’ implicit bias to alter their clini-

cal decision making in hypothetical scenarios. One

study48 found that implicit bias was related to clini-

cal decisions, the second49,51 produced mixed results,

and the third50 did not fi nd any association between

implicit bias and clinical decisions. The small size of

these studies and their reliance on hypothetical sce-

narios prevent a fi rm conclusion, but the inconsistency

of results suggests that the effect of implicit bias on

clinical decision making is not robust.

Our study was motivated by the hypothesis that

clinicians with implicit bias may communicate differ-

ently in clinical encounters with minority patients,

reducing the patients’ comfort and trust in those clini-

cians. Our results suggest that such may be the case

for black patients, but perhaps not for Latino patients.

Additional investigation will be needed to determine

why implicit bias is refl ected in the evaluations of

some groups but not others. Possible explanations

include the manner in which clinicians express bias,

the patients’ sensitivity to it, or varying expectations

and concerns.

The lack of association found in this study between

clinicians’ explicit bias and patients’ perceptions may

seem surprising—an intentionally biased clinician

ought to be viewed very poorly by minority patients.

As we and others have reported, however, clinicians

demonstrate very little explicit bias against blacks or

Latinos.26,48-52 As there is only a small probability that

minority patients would encounter an explicitly biased

clinician, there is little possibility of fi nding an associa-

tion involving that form of bias.

The fi ndings of this study are limited by several

factors. As it was an observational study, a clear assign-

ment of causality cannot be made. The lower-than-

desired patient participation rate also allows for the

possibility that response bias may have affected the

results. The type of response bias that would explain

the complex relationships obtained in this study

is, admittedly, diffi cult to imagine. The study was

restricted in scope to established primary care clini-

cians and their longer-term patients, specifi cally those

with diagnosed hypertension. In our view, this restric-

tion provides a more conservative test of the hypoth-

esis because patients are less likely to remain with

clinicians with whom they are dissatisfi ed.

In conclusion, patient-centeredness is a key compe-

tency in the training and professional development of

health care clinicians,1,53 and its importance has been

further underscored by the establishment of a national

Patient-Centered Outcomes Research Institute.54 Our

research shows that clinicians’ implicit bias may be

involved in the delivery of patient-centered care for

blacks. This fi nding supports the contention of the

Institute of Medicine18 that clinician bias may con-

tribute to health disparities, if indirectly. What might

clinicians do to avoid implicit bias in their patient

interactions? Laboratory research shows that even

though it is implicit, this form of bias is still malleable

and changes in response to specifi c alterations in situ-

ational demands and social norms.55 Helping patients

to respond to bias in a manner that helps to defl ect a

negative outcome is another path for intervention.55

For progress to be made, bias must be rendered less

implicit and unconscious, to foster real refl ection,

analysis, and change.

To read or post commentaries in response to this article, see it online at http://www.annfammed.org/content/11/1/43.

Key words: race/ethnicity; communication; prejudice; patient-centered care; healthcare disparities; primary care; practice-based research

Submitted March 25, 2012; submitted, revised, June 20, 2012; accepted July 11, 2012.

Funding support: This study was supported by grant HL088198 from the National Heart, Lung, and Blood Institute of the National Institutes of Health.

Acknowledgments: We thank the following individuals, who were compensated for their work: Natalie Wheeler, BA (Department of Psy-chology and Neuroscience, University of Colorado Boulder) for assis-tance with the patient questionnaires; Allison Ackermann (Abt SRBI Inc) for directing and managing the collection of the patient questionnaires; Anju Gupta (Institute for Health Research, Kaiser Permanente Colorado) and Brian Eckert (Denver Health) for assistance with analyzing the elec-tronic health record databases; and Stacie Daugherty, MD, MPH (School of Medicine, University of Colorado Denver) for her suggestions during the study and her helpful comments on this article.

References 1. Committee on the Quality of Health Care in America, Institute of

Medicine. Crossing the Quality Chasm: A New Health System For the 21st Century. Washington, DC: The National Academies Press; 2001.

Page 9: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

51

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

2. Beach MC, Sugarman J, Johnson RL, Arbelaez JJ, Duggan PS, Coo-per LA. Do patients treated with dignity report higher satisfaction, adherence, and receipt of preventive care? Ann Fam Med. 2005; 3(4):331-338.

3. Safran DG, Taira DA, Rogers WH, Kosinski M, Ware JE, Tarlov AR. Linking primary care performance to outcomes of care. J Fam Pract. 1998;47(3):213-220.

4. Schneider J, Kaplan SH, Greenfi eld S, Li W, Wilson IB. Better physician-patient relationships are associated with higher reported adherence to antiretroviral therapy in patients with HIV infection. J Gen Intern Med. 2004;19(11):1096-1103.

5. Wilson IB, Rogers WH, Chang H, Safran DG. Cost-related skipping of medications and other treatments among Medicare benefi ciaries between 1998 and 2000. Results of a national study. J Gen Intern Med. 2005;20(8):715-720.

6. Zolnierek KB, Dimatteo MR. Physician communication and patient adherence to treatment: a meta-analysis. Med Care. 2009;47(8): 826-834.

7. Safran DG, Montgomery JE, Chang H, Murphy J, Rogers WH. Switching doctors: predictors of voluntary disenrollment from a pri-mary physician’s practice. J Fam Pract. 2001;50(2):130-136.

8. Kaplan SH, Greenfi eld S, Ware JE Jr. Assessing the effects of physi-cian-patient interactions on the outcomes of chronic disease. Med Care. 1989;27(3 Suppl):S110-S127.

9. Kim TW, Samet JH, Cheng DM, Winter MR, Safran DG, Saitz R. Primary care quality and addiction severity: a prospective cohort study. Health Serv Res. 2007;42(2):755-772.

10. Boulware LE, Cooper LA, Ratner LE, LaVeist TA, Powe NR. Race and trust in the health care system. Public Health Rep. 2003;118(4): 358-365.

11. Collins TC, Clark JA, Petersen LA, Kressin NR. Racial differences in how patients perceive physician communication regarding cardiac testing. Med Care. 2002;40(1 Suppl):I27-I34.

12. Cooper LA, Roter DL, Johnson RL, Ford DE, Steinwachs DM, Powe NR. Patient-centered communication, ratings of care, and concor-dance of patient and physician race. Ann Intern Med. 2003;139(11): 907-915.

13. Cooper LA, Beach MC, Johnson RL, Inui TS. Delving below the sur-face. Understanding how race and ethnicity infl uence relationships in health care. J Gen Intern Med. 2006;21(Suppl 1):S21-S27.

14. Gordon HS, Street RL Jr, Sharf BF, Kelly PA, Souchek J. Racial dif-ferences in trust and lung cancer patients’ perceptions of physician communication. J Clin Oncol. 2006;24(6):904-909.

15. Johnson RL, Roter DL, Powe NR, Cooper LA. Patient race/ethnicity and quality of patient-physician communication during medical vis-its. Am J Public Health. 2004;94(12):2084-2090.

16. LaVeist TA, Nickerson KJ, Bowie JV. Attitudes about racism, medical mistrust, and satisfaction with care among black and white cardiac patients. Med Care Res Rev. 2000;57(Suppl 1):46-61.

17. Levinson W, Hudak PL, Feldman JJ, et al. “It’s not what you say ...”: racial disparities in communication between orthopedic surgeons and patients. Med Care. 2008;46(4):410-416.

18. Institute of Medicine. Smedley BD, Stith AY, Nelson AR, eds. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health-care. Washington, DC: National Academies Press; 2003.

19. van Ryn M, Fu SS. Paved with good intentions: do public health and human service clinicians contribute to racial/ethnic disparities in health? Am J Public Health. 2003;93(2):248-255.

20. Blair IV, Banaji MR. Automatic and controlled processes in stereo-type priming. J Pers Soc Psychol. 1996;70(6):1142-1163.

21. Dovidio JF, Kawakami K, Gaertner SL. Implicit and explicit prejudice and interracial interaction. J Pers Soc Psychol. 2002;82(1):62-68.

22. Gawronski B, Bodenhausen GV. Associative and propositional pro-cesses in evaluation: an integrative review of implicit and explicit attitude change. Psychol Bull. 2006;132(5):692-731.

23. Fazio RH, Jackson JR, Dunton BC, Williams CJ. Variability in auto-matic activation as an unobtrusive measure of racial attitudes: a bona fi de pipeline? J Pers Soc Psychol. 1995;69(6):1013-1027.

24. Dovidio JF, Kawakami K, Johnson C, Johnson B, Howard A. On the nature of prejudice: automatic and controlled processes. J Exp Soc Psychol. 1997;33(5):510-540.

25. McConnell AR, Leibold JM. Relations among the Implicit Associa-tion Test, discriminatory behavior, and explicit measures of racial attitudes. J Exp Soc Psychol. 2001;37(5):435-442.

26. Blair IV, Havranek EP, Price DW, et al. An assessment of biases against Latinos and blacks among primary care clinicians and com-munity members. Am J Public Health. Published online ahead of print Nov 15, 2012: e1-e7.

27. Gabow P, Eisert S, Wright R. Denver Health: a model for the inte-gration of a public hospital and community health centers. Ann Intern Med. 2003;138(2):143-149.

28. Nuzum R, McCarthy D, Gauthier A, Beck C. Denver Health: a high-performance public health system. The Commonwealth Fund, July 2007. http://www.commonwealthfund.org/Publications/Fund-Reports/2007/Jul/Denver-Health--A-High-Performance-Public-Health-Care-System.aspx. Accessed Aug 29, 2011.

29. Greenwald AG, McGhee DE, Schwartz JL. Measuring individual dif-ferences in implicit cognition: the implicit association test. J Pers Soc Psychol. 1998;74(6):1464-1480.

30. Lane KA, Banaji MR, Nosek BA, Greenwald AG. Understanding and using the Implicit Association Test: IV: What we know (so far) about the method. In: Wittenbrink B, Schwarz N, eds. Implicit Measures of Attitudes. New York, NY: Guilford Press; 2007:59-102.

31. Nosek BA, Greenwald AG, Banaji MR. Understanding and using the Implicit Association Test: II. Method variables and construct validity. Pers Soc Psychol Bull. 2005;31(2):166-180.

32. Nosek BA, Greenwald AG, Banaji MR. The Implicit Association Test at age 7: a methodological and conceptual review. In: Bargh JA, ed. Social Psychology and the Unconscious: The Automaticity of Higher Mental Processes. Florence, KY: Psychology Press; 2007:265-292.

33. Greenwald AG, Poehlman TA, Uhlmann EL, Banaji MR. Understand-ing and using the Implicit Association Test: III. Meta-analysis of predictive validity. J Pers Soc Psychol. 2009;97(1):17-41.

34. De Houwer J, Teige-Mocigemba S, Spruyt A, Moors A. Implicit measures: a normative analysis and review. Psychol Bull. 2009;135(3):347-368.

35. Blair IV, Judd CM, Havranek EP, Steiner JF. Using community data to test the discriminant validity of ethnic/racial group IATs. Zeitschrift für Psychologies. 2010;218:36-43.

36. The American National Election Studies. The ANES Guide to Public Opinion and Electoral Behavior. Ann Arbor, MI: University of Michi-gan, Center for Political Studies.

37. Smith TW, Marsden PV, Michael H, Jibum K. General Social Sur-veys, 1972-2010: Cumulative Codebook. National Data Program for the Social Sciences Series, no. 21. Chicago, IL: National Opinion Research Center; 2010.

38. Safran DG, Kosinski M, Tarlov AR, et al. The Primary Care Assess-ment Survey: tests of data quality and measurement performance. Med Care. 1998;36(5):728-739.

39. Murphy J, Chang H, Montgomery JE, Rogers WH, Safran DG. The quality of physician-patient relationships. Patients’ experiences 1996-1999. J Fam Pract. 2001;50(2):123-129.

40. Montgomery JE, Irish JT, Wilson IB, et al. Primary care experi-ences of Medicare benefi ciaries, 1998 to 2000. J Gen Intern Med. 2004;19(10):991-998.

Page 10: Clinicians’ Implicit Ethnic/Racial Bias and Perceptions of ...2012/03/25  · Ethnic/racial minorities appear to be at a disadvantage in this aspect of health care.10-17 In addition

ANNALS OF FAMILY MEDICINE ✦ WWW.ANNFAMMED.ORG ✦ VOL. 11, NO. 1 ✦ JANUARY/FEBRUARY 2013

52

CLINICIANS’ BIAS PREDIC TS C ARE PERCEPT IONS

41. Safran DG, Wilson IB, Rogers WH, Montgomery JE, Chang H. Primary care quality in the Medicare program: comparing the performance of Medicare health maintenance organizations and traditional fee-for-service Medicare. Arch Intern Med. 2002;162(7): 757-765.

42. Taira DA, Safran DG, Seto TB, et al. Do patient assessments of primary care differ by patient ethnicity? Health Serv Res. 2001;36(6 Pt 1):1059-1071.

43. Morales LS, Cunningham WE, Brown JA, Liu H, Hays RD. Are Lati-nos less satisfi ed with communication by health care providers? J Gen Intern Med. 1999;14(7):409-417.

44. Weech-Maldonado R, Morales LS, Elliott M, Spritzer K, Marshall G, Hays RD. Race/ethnicity, language, and patients’ assess-ments of care in Medicaid managed care. Health Serv Res. 2003;38(3):789-808.

45. Stewart AL, Nápoles-Springer AM, Gregorich SE, Santoyo-Olsson J. Interpersonal processes of care survey: patient-reported measures for diverse groups. Health Serv Res. 2007;42(3 Pt 1):1235-1256.

46. Penner LA, Dovidio JF, West TV, et al. Aversive racism and medi-cal interactions with black patients: a fi eld study. J Exp Soc Psychol. 2010;46(2):436-440.

47. Cooper LA, Roter DL, Carson KA, et al. The associations of clini-cians’ implicit attitudes about race with medical visit communica-tion and patient ratings of interpersonal care. Am J Public Health. 2012;102(5):979-987.

48. Green AR, Carney DR, Pallin DJ, et al. Implicit bias among physi-cians and its prediction of thrombolysis decisions for black and white patients. J Gen Intern Med. 2007;22(9):1231-1238.

49. Sabin JA, Rivara FP, Greenwald AG. Physician implicit attitudes and stereotypes about race and quality of medical care. Med Care. 2008;46(7):678-685.

50. Haider AH, Sexton J, Sriram N, et al. Association of unconscious race and social class bias with vignette-based clinical assessments by medical students. JAMA. 2011;306(9):942-951.

51. Sabin JA, Greenwald AG. The infl uence of implicit bias on treatment recommendations for 4 common pediatric conditions: pain, urinary tract infection, attention defi cit hyperactivity disorder, and asthma. Am J Public Health. 2012;102(5):988-995.

52. Sabin JA, Nosek BA, Greenwald AG, Rivara FP. Physicians’ implicit and explicit attitudes about race by MD race, ethnicity, and gender. J Health Care Poor Underserved. 2009;20(3):896-913.

53. Greiner AC, Knebel E, eds. The Core Competencies Needed for Health Care Professionals. Health Professions Education: A Bridge to Quality. Washington, DC: National Academy Press; 2003.

54. Patient-Centered Outcomes Research Institute. 2012. http://www.pcori.org. Accessed Aug 1, 2011.

55. Blair IV, Steiner JF, Havranek EP. Unconscious (implicit) bias and health disparities: where do we go from here? Permanente J. 2011; 15(2):71-78.

CHANGE-OF-ADDRESS FORMPlease complete this form and mail to the following address or fax to Annals Circulation at 913-906-6080:

Annals of Family Medicine, Circulation Department, 11400 Tomahawk Creek Pkwy, Leawood, KS 66211-2680

Check if member of sponsoring organization: ❏ AAFP ❏ ABFM ❏ STFM ❏ ADFM

❏ AFMRD ❏ NAPCRG ❏ CFPC

ID number from label on your journal cover __ __ __ __ __ __ __

OLD Information (Please print.)

Name

Company (if applicable)

Address (Street plus Apt or Ste)

City State

Country Postal Code (9-digit ZIP for US)

Telephone Fax

E-Mail

NEW Information (Please print.)

Name

Company (if applicable)

Address (Street plus Apt or Ste)

City State

Country Postal Code (9-digit ZIP for US)

Telephone Fax

E-Mail